3D model reconstruction based on plantar image's feature segmentation

Abstract

The 3D reconstruction of sole based on laser scanning is costly and time-consuming, the 3D reconstruction based on structured light is susceptible to ambient light. A 3D reconstruction method of sole based on the plantar image's feature segmentation is proposed. The sole is segmented into two regions by the morphological features in the image. One is the arch part touching the glass, another is the un-touching camber part. And an experiment is designed to summarize the law between the plantar height and gray value in each region. Then we reconstruct the two regions' shapes separately on the basis of the regulation. Finally, we implement the plantar reconstruction by merging the two point cloud models. Experimental evaluation demonstrates that the system can do the classification and reconstruction effectively.